Siemens Digital Industries Software explained its "AI-native EDA" strategy, which integrates AI into semiconductor and Printed Circuit Board (PCB) design, as well as its progress toward this goal.
Product LaunchesSiemens Digital Industries SoftwareSiemens EDA
Siemens EDA Promotes Semiconductor Design Automation via Autonomous AI Agents
This article is a translation. Read the Japanese original
The AI agents developed by the company function as "long-running agents" capable of calling and executing multiple EDA tools based on goals set by designers. Even if problems arise during processing, they can identify the causes via inference, correct settings, and continue working until the design objectives are met.
Through this initiative, the company aims to create an environment where repetitive tasks faced by designers, such as revision corrections and error analysis, are shifted to AI, allowing designers to focus on architectural considerations. Siemens EDA stated that it achieves productivity improvements of 10 to 50 times depending on the design process.
To ensure the reliability of AI-generated results, the company is also working on building "self-verifying AI agents" that continuously verify results using physics-based EDA engines. Additionally, the company has adopted a "Bring Your Own Model" policy that does not depend on a specific model, allowing for configurations that utilize existing general-purpose LLMs (Large Language Models) such as NVIDIA's Nemotron.
Sources
- The Era of Utilizing AI Agents in Semiconductor Design Has Arrived: Siemens EDA Outlines the Division of Roles Between Humans and AI （Editorial Select: AI-driven foundation design (2026-09-16), 2026-09-07）